I've never looked at IXGLOGR address space but my guess is that IXGLOGR would have multiple tasks (TCB's) running at the same time if there are multiple logs active. As for batch running slower at night after you went from 1 CPU to 4, that doesn't make sense unless other things changed. SRM dispatching? More workload? Competition for resources such as disk usage? LPARs and LPAR configurations? CPUs dedicated to an LPAR versus shared CPUs. As for 20 years later, an address space has always been able to support multiple dispatchable tasks. I've worked on address spaces with over a hundred tasks with over 10 dispatchable tasks at one time. All could be active as long as enough CPUs are available. On Wednesday, July 12, 2023 at 03:27:32 PM PDT, Jason Cai <[email protected]> wrote: Dear all,
I have a question about CPU usage on z/OS that I hope you can help me with. Many years ago, we increased the CPU of a mainframe from one to four, but the batch jobs at night became slower. The explanation at that time was that a batch job could only use one CPU, and the single CPU performance decreased because of the increase to four CPUs. I accepted that explanation at that time. Now, twenty years later, is z/OS still the same that an address space like IXGLOGR can only use one CPU at most? Thank you for your time and attention. I look forward to hearing from you soon. Jason Cai ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
